Notes from DGAS Sensitivity Testing Meeting, May 20, 1997

Columbia Basin Research hosts this web page for archival purposes only. This project/page/resource is no longer active.

Attending: Tom Carlson, Rock Peters, Jim Anderson, Pam Shaw, Larry Fidler, Chris Pinney, Marshall Richmond, Kim Fodrea


Development and application of a simplified numerical testbed to test the sensitivity of the developed numerical models to the effects of different DGAS alternatives as they ultimately affect fish mortality and dissolved gas levels in the river system. The general ideas for the testbed and analysis are presented in Appendix A. The following sketch shows the general concept of building a testbed consisting of a simplified project and reservoir system.

Figure - Sketch showing the general layout of the test reservior and components


Tom Carlson stated that the time has come to bring together all of the efforts and research put forth by the Corps under the Gas Abatement Study and present what has been learned to the region. Performing a sensitivity analysis for the hydrodynamic and mortality models is the start of this process. In particular, the Corps needs a tool to assess the biological benefits of reducing TDG even by just a few percent. It is also hoped that by doing the sensitivity analysis any future research needs can be identified in time to be put forth on the coming year's budget.

Jim presented a list of typical issues important to a manager:

  1. Reduction of physical injury in the tailrace zone
  2. Retention in stilling basin/tailrace zone - how sensitive is mortality to this?
  3. Tailrace gas concentrations and distribution
  4. Identification of areas and timing of elevated gas concentration in the reservoir
  5. Vertical distribution of fish
  6. How could spill be used to reduce predation.

In order to complete the sensitivity analysis the following things must be identified: what measure should be used to evaluate benefits from the alternatives and what information is needed for this measure, what are the relevant time scales for each of the numerical models, what set of hypotheses will be tested under the sensitivity analysis.

Time and spatial scale of the hydrodynamic model would be in the range of 5 minutes and 100 feet. Model needs to address 3 mortality mechanisms: TDG/GBT, predation, and physical injury.

Output of this effort is a framework to complete the full alternatives analysis for DGAS. This effort will also help direct future research and has a high priority to be completed soon, as next years research planning cycle is underway.


The tasks were broken into three phases.

  • Phase I - Development of the physical and fish behavior components. Generation of initial exposure logs.
  • Phase II - Addition of gas bubble trauma computations to assess mortality associated with the exposure logs
  • Phase III - Sensitivity Analysis for a wide range of alternatives

Each phase has different components that need to be developed and applied. This table shows the components, which phase they fit into, and who the lead(s) are for that element.

Physical Biological Stilling Basin Zone Issues
I - 2D Hydrodynamic Model (MR) I - Fish Behavior Rules (JA,PS) II - Fish Retention Time (TC)
I - 2D Water Quality Model (MR) II - Broken Stick GBT Model as used in CRiSP.1.5 (JA,PS) II - Physical Damage Estimates (TC)
I - 2D Fish Particle Tracking Module (MR) II - "Static" GBT Model (LF) II - Gas Concentration Measurements(Carroll, et al.)
I - Generation of Exposure Logs (MR) II - "Dynamic" GBT Model (LF) II - Estimates of Gas Production (Wilhelms, et al.)
II - Integration with GBT Models II - "Vitality" GBT Model (JA) --

Schedule for Sensitivity Testing

The group developed this schedule for completion of the sensitivity test task.

Item Date
Complete Phase I Development July 1, 1997
Complete Initial Exposure Log July 30, 1997
Complete Phase II Development/Integration
Intermediate Deliverable Due
August 15, 1997
Complete Sensitivity Analysis September 30, 1997
Complete Documentation of Models and Alternatives March 1998

Schedule for 60% Report Completion

Kim presented the following schedule for completing the 60% report. The dates material needs to be submitted to Kim are listed. Chris, Rock, and Tom need to define the dates material needs to be sent to them prior to Kim's submittal dates.

Item Draft to Rock, et al. Draft Back from Rock, et al. Draft to Kim Draft back from Kim
1st Draft - - October 15, 1997 October 22, 1997
2nd Draft - - December 11, 1997 December 18, 1997
3rd Draft - - January 22, 1998 February 4, 1998
Final Revisions - - February 12, 1998 Point of No Return
Report to Printer February 27, 1998

Appendix A - Strawman Sensitivity Plan

The following is the set of notes that Pam Shaw and Marshall Richmond put together prior to the meeting.


Test mortality sensitivity to various DGAS alternatives and modeling uncertainties. Also assess modeling issues such as computation time, storage requirements, etc.


Apply flow, gas, and biology models to a simple box reservoir representing a reach starting at the spillway of one dam and ending at the forebay of the next project. The hydrodynamics will be computed using a general two-dimensional depth-averaged model. Corresponding two-dimensional gas transport will be simulated for each flow with gas production at the hypothetical project specified using the WES equations. The biology will be modeled using an individual based model (IBM) that will using particle tracking methods to compute exposure logs for a large number of individuals (10,000 is the target). These exposure logs will be fed into the different mortality models to estimate the resultant moralities for the different scenarios.


The following are the main components for which sensitivity tests need to be done:

  1. DGAS alternatives
  2. Physical scenarios
  3. Model Parameters
  4. Fish Behavior
  5. Mortality Models
  6. Analysis and Presentation


  1. Assign alternative - no action, raise stilling basin
  2. Compute flow and gas fields for the alternative or for different powerhouse/spill discharges for a given alternative. This step may generate several different flow/gas realizations for later use in the IBM. Save these for use in step 3.
  3. Assign fish behavior - depth distribution, downstream migration velocity, lateral movement, and depth redistribution after passage, diel effects.
  4. Compute exposure logs for the specified flow/gas realization and fish behavior rules. Save these for use in step 5 and 6.
  5. Compute fish mortality using the different mortality models: existing CRiSP method, vitality model, and the dynamic gas bubble model.
  6. Assess Sensitivity for the alternative - % mortality in reach over a season, % mortality over a day, % mortality in specific zone such as the stilling basin, others? What are the appropriate measures?



  1. Alternatives: which alternatives will be tested.
  2. Physical scenarios: flow levels, spill regimes, spill patterns, etc.
  3. Gas Production: the effect of each alternative on gas production
    ( New gas functions? Apply a % decrease to existing gas production?)
  4. Fish Behavior: Different behavior rules for the fish: vertical movement, lateral movement, depth distributions, diel patterns. Possibly multiple scenarios corresponding to different fish behavior hypotheses.
  5. Gas Model Parameters: Determine sensitivity to parameter choice under the current state of the system and settle on mixing and dissipation parameters for the hypothetical reach.
  6. Mortality Models: Resolution of gas model output that each mortality model requires
  7. Sensitivity Measures: Identify appropriate measures of TDG mortality under the different alternatives.